Marc Wichern

Biogas from grass silage – Measurements and modeling with ADM1

Mono fermentation of grass silage without the addition of manure was performed over a period of 345days under mesophilic conditions (38 degrees C). A simulation study based on the IWA Anaerobic Digestion Model No. 1 (ADM1) was done in order to show its applicability to lignocellulosic biomass. Therefore, the influent was fractioned by established fodder analysis (Weender analysis and van Soest method). ADM1 was modified with a separate compound of inert decay products similar to the approach of Activated Sludge Model No. 1 (ASM1). Furthermore, a function, which described the influence of solids on the process of hydrolysis, has been integrated to reproduce reliable ammonium concentrations. The model was calibrated by using the modified Nash-Sutcliffe coefficient to evaluate simulation quality. It was possible to fit observed data by changing only hydrogen inhibition constants and the maximum acetate uptake rate. The extended ADM1 model showed good agreement with measurements and was suitable for modeling anaerobic digestion of grass silage.

Mono fermentation of grass silage by means of loop reactors

A loop reactor was operated for mono fermentation of grass silage without manure addition under mesophilic conditions (38 degrees C). An averaged specific biogas production of 0.50 m(N)(3) per kg volatile solids (VS) with a methane concentration of 52% at an organic loading rate of up to 3.5 kg(VS)/(m(3) d) was obtained. The retention time varied from 440 days at 1.0 kg(VS)/(m(3) d) to 50 days at 3.5 kg(VS)/(m(3) d). The degradation level was more than 60% based on VS and 75% based on COD. The first-order hydrolysis rate constant of the process was estimated to be 0.6 d(-1). Despite the relative high ammonium concentration of up to 4 g/l, the system worked stable for an operation period of 310 days. In particular the TS content in the fermenter was found to be a key parameter and should not exceed 12% in order to avoid instabilities.

Monofermentation of grass silage under mesophilic conditions: Measurements and mathematical modeling with ADM 1

In this paper experimental data from grass fermentation and simulation results with the Anaerobic Digestion Model (ADM) No. 1 are described. Two laboratory reactors were operated under mesophilic conditions with volumetric loading rates in between 0.3 and 2.5 kg(VS)/(m(3) x d). Two different kinds of grass silage were used as substrates, resulting in an average specific biogas production of 600 L/kg(VS). The ADM 1 was calibrated both manually and with the help of a Genetic Algorithm in Matlab/Simulink. Results from calibration indicate that the NH3 inhibition constant used to model the inhibition of acetate uptake is three to five times higher compared with digested activated sludge. The hydrogen inhibition constants applied for propionate and valerate/butyrate uptake are around two orders of magnitude lower than for sludge digestion.

Microbiological fermentation of lignocellulosic biomass: current state and prospects of mathematical modeling

The anaerobic fermentation process has achieved growing importance in practice in recent years. Anaerobic fermentation is especially valuable because its end product is methane, a renewable energy source. While the use of renewable energy sources has accelerated substantially in recent years, their potential has not yet been sufficiently exploited. This is especially true for biogas technology. Biogas is created in a multistage process in which different microorganisms use the energy stored in carbohydrates, fats, and proteins for their metabolism. In order to produce biogas, any organic substrate that is microbiologically accessible can be used. The microbiological process in itself is extremely complex and still requires substantial research in order to be fully understood. Technical facilities for the production of biogas are thus generally scaled in a purely empirical manner. The efficiency of the process, therefore, corresponds to the optimum only in the rarest cases. An optimal production of biogas, as well as a stable plant operation requires detailed knowledge of the biochemical processes in the fermenter. The use of mathematical models can help to achieve the necessary deeper understanding of the process. This paper reviews both the history of model development and current state of the art in modeling anaerobic digestion processes.

The impact of sunlight on inactivation of indicator microorganisms both in river water and benthic biofilms

A detailed knowledge on decay or inactivation kinetics of faecal indicator microorganism in rivers is essential for control of bathing water quality. Both reliable inactivation rate coefficients for such kinetics and the knowledge on pathogen accumulation in benthic biofilms are needed for the assessment of river self purification capability and for being able to make decisions for an optimized water management. Therefore, the inactivation kinetics for main indicator microorganisms like faecal coliforms and intestinal Enterococci in water and on stones of the river Isar (Germany) were measured at artificial sunlight radiation. A flume was built, containing Isar water and sediment. By adding a pulse of pre-treated sewage water, the subsequently change of indicator bacteria levels in the bulk and biofilm phase is measured. Bacterial inactivation was mainly dependent on sunlight intensity. Mean inactivation rate coefficients in the experimental set up with river water obtained for a radiation intensity comparable to average midday sunlight in June in 50 degrees north latitude (I((290-390nm))=40.0W/m(2)) were found to be 21.4d(-1) for faecal coliforms and 20.0d(-1) for intestinal Enterococci. For a radiation intensity conform with the annual mean radiation in Germany (I((290-390nm))=8.0W/m(2)) inactivation rate coefficients were 12.7d(-1) for faecal coliforms and 9.3d(-1) for intestinal Enterococci. A measurement without direct artificial sunlight (I((290-390nm))=0.08W/m(2)) yields inactivation rate coefficients of 3.4d(-1) for faecal coliforms and 1.7d(-1) for intestinal Enterococci, which were similar to those found in wastewater treatment ponds. Due to re-growth and better environmental conditions the concentrations of faecal coliforms and intestinal Enterococci within the biofilm were 10(2) to 10(4) fold higher compared to the bulk water depending on the exposure time.

An integrated 45 L pilot microbial fuel cell system at a full-scale wastewater treatment plant

A 45-L pilot MFC system, consisting of four single-chamber membraneless MFCs, was integrated into a full-scale wastewater treatment plant (WWTP) and operated under practical conditions with the effluent of the primary clarifier for nine months to identify an optimal operational strategy for stable power output and maximum substrate based energy recovery (Normalized Energy Recovery, NER). Best results with the MFC were obtained at a hydraulic retention time of 22h with COD, TSS and nitrogen removal of 24%, 40% and 28%, respectively. Mean NER of 0.36kWhel/kgCOD,deg and coulombic efficiency of 24.8% were reached. Experimental results were used to set up the first described energy balance for a whole WWTP with an integrated MFC system. Energetic calculations of the model WWTP showed that energy savings due to reduced excess sludge production and energy gain of the MFC are significantly higher than the loss of energy due to reduced biogas production.

Investigations and mathematical simulation on decentralized anaerobic treatment of agricultural substrate from livestock farming

Anaerobic processes are widely used for treatment of both municipal and industrial wastewater as well as agricultural substrates. In contrast to the aerobic methods, they are frequently more cost-efficient, they have a lower surplus sludge production, and the reactors can be run with higher volumetric loads and thus smaller volumes. In the paper presented both experimental data and the application of the Anaerobic Digestion Model No. 1 for agricultural substrate from livestock farming will be described. A 3,500 L reactor with mesophilic operation and loaded with cattle manure was examined with respect to its COD degradation, gas production, and gas composition. Results revealed a reduction of 30-35% COD and a biogas production of 287 L(Biogas)/kg(VS) when operated with a specific loading rate of 3.6 kg(VS)/(m(3).d).After calibration of the ADM 1, which was based predominantly on the acetate uptake rate (k(ac.m)=3.6 g/(g.d)), the disintegration constant (k(Dis)=0.05 d(-1)) and the exact determination of the influent COD fractions contained in the agricultural substrate, it was possible to simulate the measured data of the plant in excellent quality. For future application of the ADM 1 as part of control strategies a sensitivity analysis was carried out. The analysis based on the SVM slope technique has been done to identify highly sensitive biochemical parameters. These are, amongst others, the acetate uptake rate, the disintegration constant, the biomass decay rates and the half saturation constant for ammonia inhibition. Sensitivity analysis of the inflow COD fractions (proteins, carbohydrates, lipids and inert) showed the necessity of detailed measurements for the prediction of the gas flow and composition as well as for prognosis of inhibitions in the anaerobic process. For cattle manure especially the fractions of inert material and carbohydrates should be observed carefully. Due to the high content of NH(4)-N in manure the protein fraction is not as sensitive as the two mentioned above.

Experimental results and mathematical modelling of an autotrophic and heterotrophic biofilm in a sand filter treating landfill leachate and municipal wastewater.

A better understanding of wastewater treatment with soil filters is important to optimise plant operation and reduce the risk of clogging. The article presents results of a treatment concept which uses a combination of SBR and vertical-flow sand filter technology. The SBR was mainly used for denitrification and sedimentation of substances in particulate form. Efficient nitrification was achieved by the planted sand filter. Degradation rates of 10gNH(4)-N/(m(2)xd) were measured for periods with peak loadings. The two-dimensional dynamic model reproduces the biofilm growth and decay of heterotrophic and autotrophic biomass. It is capable of describing the clogging of the sand filter by combining a biochemical and a geometric model. After calibration, the model was used for the calculation of maximum nitrogen degradation performances. Maximum degradation rates of 12gNH(4)-N/(m(2)xd) can be achieved if the COD/TKN ratio is reduced before to a level lower than that of municipal wastewater. The COD was further degraded in the filter than we expected comparing it with activated sludge plants. Within the soil filter a biofilm thickness of up to 110microm is simulated depending on the embankment of gravel and grains of sand. Sensitivity analysis of model parameters showed the high impact of the maximum autotrophic growth rate, the autotrophic yield, the diffusion coefficient for oxygen and the number of contact points of the single grains of sand.

Determination of Methanogenic Pathways through Carbon Isotope (δ13C) Analysis for the Two-Stage Anaerobic Digestion of High-Solids Substrates

This study used carbon isotope (δ(13)C)-based calculations to quantify the specific methanogenic pathways in a two-stage experimental biogas plant composed of three thermophilic leach bed reactors (51-56 °C) followed by a mesophilic (36.5 °C) anaerobic filter. Despite the continuous dominance of the acetoclastic Methanosaeta in the anaerobic filter, the methane (CH4) fraction derived from carbon dioxide reduction (CO2), fmc, varied significantly over the investigation period of 200 days. At organic loading rates (OLRs) below 6.0 gCOD L(-1) d(-1), the average fmc value was 33%, whereas at higher OLRs, with a maximum level of 17.0 gCOD L(-1) d(-1), the fmc values reached 47%. The experiments allowed for a clear differentiation of the isotope fractionation related to the formation and consumption of acetate in both stages of the plant. Our data indicate constant carbon isotope fractionation for acetate formation at different OLRs within the thermophilic leach bed reactors as well as a negligible contribution of homoacetogenesis. These results present the first quantification of methanogenic pathway (fmc values) dynamics for a continually operated mesophilic bioreactor and highlight the enormous potential of δ(13)C analysis for a more comprehensive understanding of the anaerobic degradation processes in CH4-producing biogas plants.

Optimizing sequencing batch reactor (SBR) reactor operation for treatment of dairy wastewater with aerobic granular sludge

The biological wastewater treatment using aerobic granular sludge is a new and very promising method, which is predominantly used in SBR reactors which have higher volumetric conversion rates than methods with flocculent sludge. With suitable reactor operation, flocculent biomass will accumulate into globular aggregates, due to the creation of increased substrate gradients and high shearing power degrees. In the research project described in this paper dairy wastewater with a high particle load was treated with aerobic granular sludge in an SBR reactor. A dynamic mathematical model was developed describing COD and nitrogen removal as well as typical biofilm processes such as diffusion or substrate limitation in greater detail. The calibrated model was excellently able to reproduce the measuring data despite of strongly varying wastewater composition. In this paper scenario calculations with a calibrated biokinetic model were executed to evaluate the effect of different operation strategies for the granular SBR. Modeling results showed that the granules with an average diameter of 2.5 mm had an aerobic layer in between 65-95 microm. Density of the granules was 40 kgVSS/m3. Results revealed amongst others optimal operation conditions for nitrogen removal with oxygen concentrations below 5 gO2/m3. Lower oxygen concentrations led to thinner aerobic but thicker anoxic granular layers with higher nitrate removal efficiencies. Total SBR-cycle times should be in between 360-480 minutes. Reduction of the cycle time from 480 to 360 minutes with a 50% higher throughput resulted in an increase of peak nitrogen effluent concentrations by 40%. Considering biochemical processes the volumetric loading rate for dairy wastewater should be higher than 4.5 kgCOD/(m3*d). Higher COD input load with a COD-based volumetric loading rate of 9.0 kgCOD/(m3*d) nearly led to complete nitrogen removal. Under different operational conditions average nitrification rates up to 5 gNH/(m3*h) and denitrification rates up to 3.7 gNO/(m3*h) were achieved.